Download presentation

Presentation is loading. Please wait.

Published byYuliana Dugdale Modified over 2 years ago

1
One-Way BG ANOVA Andrew Ainsworth Psy 420

2
Topics Analysis with more than 2 levels Deviation, Computation, Regression, Unequal Samples Specific Comparisons Trend Analysis, Planned comparisons, Post- Hoc Adjustments Effect Size Measures Eta Squared, Omega Squared, Cohen’s d Power and Sample Size Estimates

3
Deviation Approach When the n’s are not equal

4
Analysis - Traditional The traditional analysis is the same

5
Analysis - Traditional Traditional Analysis – Unequal Samples

6
Unequal N and DFs

7
Analysis - Regression In order to perform a complete analysis of variance through regression you need to cover all of the between groups variance To do this you need to: Create k – 1 dichotomous predictors (Xs) Make sure the predictors don’t overlap

8
Analysis – Regression

9
One of the easiest ways to ensure that the comps do not overlap is to make sure they are orthogonal Orthogonal (independence) The sum of each comparison equals zero The sum of each cross-product of predictors equals zero

10
Analysis – Regression

11
Analysis - Regression

13
Analysis – Regression

14
Formulas

15
Analysis – Regression Formulas

16
Analysis – Regression Example

17
Analysis - Regression Example

18
Analysis - Regression Example

19
Analysis - Regression Example

20
Analysis - Regression Example F crit (2,12) = 3.88, since 12.253 is greater than 3.88 you reject the null hypothesis. There is evidence that drug type can predict level of anxiety

21
Analysis - Regression Example

22
Analysis - Regression SPSS

23
Analysis - Regression SPSS

24
Analysis - Regression SPSS

25
Specific Comparisons F-test for Comparisons n = number of subjects in each group = squared sum of the weighted means = sum of the squared coefficients MS S/A = mean square error from overall ANOVA

26
Specific Comparisons If each group has a different sample size…

27
Specific Comparisons Example

28
Specific Comparisons Trend Analysis If you have ordered groups (e.g. they differ in amount of Milligrams given; 5, 10, 15, 20) You often will want to know whether there is a consistent trend across the ordered groups (e.g. linear trend) Trend analysis comes in handy too because there are orthogonal weights already worked out depending on the number of groups (pg. 703)

29
Specific Comparisons Different types of trend and coefficients for 4 groups

30
Specific Comparisons Mixtures of Linear and Quadratic Trend

31
Specific Comparisons Planned comparisons - if the comparisons are planned than you test them without any correction Each F-test for the comparison is treated like any other F-test You look up an F-critical value in a table with df comp and df error.

32
Specific Comparisons Example – if the comparisons are planned than you test them without any correction… F x1, since 11.8 is larger than 4.75 there is evidence that the subjects in the control group had higher anxiety than the treatment groups F x2, since 12.75 is larger than 4.75 there is evidence that subjects in the Scruital group reporter lower anxiety than the Ativan group

33
Specific Comparisons Post hoc adjustments Scheffé This is used for complex comparisons, and is conservative Calculate F comp as usual F S = (a – 1)F C where F S is the new critical value a – 1 is the number of groups minus 1 F C is the original critical value

34
Specific Comparisons Post hoc adjustments Scheffé – Example F X1 = 11.8 F S = (3 – 1) * 4.75 = 9.5 Even with a post hoc adjustment the difference between the control group and the two treatment groups is still significant

35
Specific Comparisons Post hoc adjustments Tukey’s Honestly Significant Difference (HSD) or Studentized Range Statistic For all pairwise tests, no pooled or averaged means F comp is the same, q T is a tabled value on pgs. 699-700

36
Specific Comparisons Post hoc adjustments Tukey’s Honestly Significant Difference (HSD) or Studentized Range Statistic Or if you have many pairs to test you can calculate a significant mean difference based on the HSD, where q T is the same as before, when unequal samples

37
Specific Comparisons Post hoc adjustments Tukey’s – example Since 12.74 is greater than 7.11, the differences between the two treatment groups is still significant after the post hoc adjustment

38
Specific Comparisons Post hoc adjustments Tukey’s – example Or you calculate: This means that any mean difference above 1.79 is significant according to the HSD adjustment 7.2 – 4.8 = 2.4, since 2.4 is larger than 1.79…

39
Effect Size A significant effect depends: Size of the mean differences (effect) Size of the error variance Degrees of freedom Practical Significance Is the effect useful? Meaningful? Does the effect have any real utility?

40
Effect Size Raw Effect size – Just looking at the raw difference between the groups Can be illustrated as the largest group difference or smallest (depending) Can’t be compared across samples or experiments

41
Effect Size Standardized Effect Size Expresses raw mean differences in standard deviation units Usually referred to as Cohen’s d

42
Effect Size Standardized Effect Size Cohen established effect size categories.2 = small effect.5 = moderate effect.8 = large effect

43
Effect Size Percent of Overlap There are many effect size measures that indicate the amount of total variance that is accounted for by the effect

44
Effect Size Percent of Overlap Eta Squared simply a descriptive statistic Often overestimates the degree of overlap in the population

45
Effect Size Omega Squared This is a better estimate of the percent of overlap in the population Corrects for the size of error and the number of groups

46
Effect Size Example

47
Effect Size For comparisons You can think of this in two different ways SS comp = the numerator of the F comp

48
Effect Size For comparisons - Example

49
Power and Sample Size Designing powerful studies Select levels of the IV that are very different (increase the effect size) Use a more liberal α level Reduce error variability Compute the sample size necessary for adequate power

50
Power and Sample Size Estimating Sample size There are many computer programs that can compute sample size for you (PC-Size, G-power, etc.) You can also calculate it by hand: Where 2 = estimated MS S/A = desired difference Z α-1 = Z value associated with 1 - α z -1 = Z value associated with 1 -

51
Power and Sample Size Estimating Sample size – example For overall ANOVA with alpha =.05 and power =.80 (values in table on page 113) Use the largest mean difference Roughly 2 subjects per group For all differences significant Roughly 28 subjects per group

Similar presentations

OK

Chapter 13 For Explaining Psychological Statistics, 4th ed. by B. Cohen 1 Chapter 13: Multiple Comparisons Experimentwise Alpha (α EW ) –The probability.

Chapter 13 For Explaining Psychological Statistics, 4th ed. by B. Cohen 1 Chapter 13: Multiple Comparisons Experimentwise Alpha (α EW ) –The probability.

© 2018 SlidePlayer.com Inc.

All rights reserved.

Ads by Google

Ppt on world environment day 2016 Ppt online shopping project in html Ppt on computer languages systems Ppt on bodybuilding supplements Ppt on word association test in psychology Ppt on inhabiting other planets in the universe Ppt on environmental issues in india Ppt on power generation using footsteps prayer Ppt on water pollution free download Genetics for kids ppt on batteries